karimalami7 / mssd Goto Github PK

Dear Doctor Alami Karim,

I hope this message finds you in good health and high spirits. My name is Walid Khames, and I am a student currently focused on the parallelization of database queries. While exploring various resources, I came across your C++ MSSD implementation on GitHub, which has proven to be invaluable in understanding the benchmarking of parallel Skyline algorithms.

I am particularly interested in investigating memory usage for Skyline algorithms, and I am reaching out to seek your expertise on the matter. In the context of your implementation, if you were to measure memory usage for either the NSct, SkyBench, or DBsky algorithms, what tools or methodologies would you recommend? My goal is to conduct a comparative analysis of memory usage between sequential and parallel skyline executions, with a specific emphasis on evaluating the potential impact of parallelizing the skyline algorithm on memory consumption.

Your insights on this matter would greatly contribute to my research, and I am eager to incorporate any recommendations or ideas you may have. I understand that your time is valuable, and I sincerely appreciate your consideration of my inquiry.

I asked Pr Sean Chester who wrote Skybench implementation and he suggested this: Regarding the C++ (i.e., multi-core, not CUDA GPU) code, I generally measure memory consumption as the "high watermark" in /proc/[pid]/status using Linux. It's unfortunately a bit manual, as you have to read that file just before the process terminates, but one can add a long sleep and a notification to make that easier.

Thank you for your time, and I look forward to hearing from you.
walid khames